Enjo | Troopr Labs
Engage Visitors Instantly, Qualify Leads Automatically, and Convert Faster.
Designing a conversational AI agent that helps businesses automate support, retrieve knowledge, and assist users directly from their website.
The Rise of AI Support Agents
Designing an AI Agent for Conversational Workflows
Customer support is rapidly shifting from traditional chatbots to AI-powered agents that can understand context, retrieve information, and assist users autonomously.
While most companies still rely on rule-based bots, modern AI agents can:
• Answer complex queries
• Retrieve knowledge from documentation
• Assist users in real time
• Escalate to humans when needed
However, deploying these agents on websites is often complex, rigid, and poorly designed.
This project explores how businesses can deploy an AI agent through a lightweight widget that integrates seamlessly into their website.
AI Agent
Improving the Experience of an Embeddable AI Agent Widget
AI agents are increasingly becoming the first touchpoint between businesses and their customers. Many companies deploy AI-powered widgets on their websites to automate support, answer questions, and guide users.
However, while these systems are powerful, their interface and interaction design often create friction for users.
This project focuses on improving the experience of an existing AI agent widget by making the interaction clearer, more intuitive, and easier for users to engage with.
The goal was not to redesign the entire product, but to enhance the usability and interaction quality of the AI agent interface.
Research
Understanding the Existing Experience
The AI agent widget allows website visitors to interact with an intelligent assistant capable of answering questions and retrieving information from company knowledge sources.
Users can open the widget, ask questions, and receive AI-generated responses.
While the functionality was powerful, the experience had several usability challenges that affected how easily users could interact with the assistant.
Some of the key areas of improvement included:
• clarity of the widget entry point
• guidance for first-time users
• conversation readability
• discoverability of AI capabilities
These issues created an opportunity to refine the interface and improve the overall experience.
Ideation
Identifying UX Opportunities
After reviewing the existing interface, I focused on identifying areas where the interaction could be improved without changing the underlying system.
Three key opportunities emerged.
Improving Discoverability:
Users should immediately understand that the widget is an AI assistant that can help them find information or solve problems.
Reducing Interaction Friction:
First-time users often struggle to know what they should ask the AI agent. The experience needed better guidance.
Enhancing Conversation Clarity:
AI responses should be structured in a way that is easy to read and easy to follow.These opportunities guided the improvements made in the interface.
Design Improvements
This project focused on designing the interface layer of an AI agent system, translating complex AI capabilities into a simple and usable experience for both businesses and end users.
Tailored Onboarding
How can we improve the onboarding experience?
Getting started with Enjo is now faster, simpler, and more personalized. The new onboarding experience adapts to your goals across key use cases like Customer Support, Sales, and HR, guiding you through only what’s relevant.
With quick verification, smart default settings, and instant knowledge setup, teams can go from signup to value in minutes. Every new workspace helps users launch AI agents faster and achieve meaningful results right away.
Enhanced Navigation
How can we improve users' navigation action with less confusion switching between agents?
We’ve enhanced the Enjo experience with a redesigned user interface featuring a streamlined two-panel navigation system. This update makes it easier to access core modules like Agent Studio, Enjo Inbox, and Insights—reducing clicks and improving day-to-day usability.
By consolidating essential functions into a streamlined view, this upgrade reduces friction in everyday usage and makes the entire product easier to adopt and operate at scale.
Guided Previews
How to improve new users' knowledge for the particular module in the tool?
To address this, the interface introduces suggested prompts that help users start conversations quickly.
These prompts reduce friction and make the AI assistant feel more approachable.
Enhanced Enjo Webchat
How to improve webchat customization and deployment experience with a preview?
The updated Enjo Webchat brings a smoother, more intuitive setup process for designing, testing, and launching customer chat experiences. With an improved layout and real-time preview, teams can instantly see and validate design changes. The built-in testing tools allow you to perfect chat interactions before going live, enabling faster deployment of on-brand, high-quality customer experiences.
Final Experience
The improved AI agent widget provides a more intuitive conversational experience by focusing on clarity, guidance, and readability.
Key improvements include:
• clearer widget entry point
• guided prompts for first interactions
• improved message structure
• simplified conversation layout
These refinements help users interact with the AI assistant more naturally while maintaining the existing system capabilities.
Reflection
Working on this experience improvement highlighted how important interaction design is for AI-powered interfaces.
Even when the underlying technology is powerful, the experience can still feel confusing without thoughtful UX design.
Small improvements in guidance, readability, and interaction patterns can significantly improve how users engage with AI systems.
This project reinforced the importance of designing AI interfaces that feel clear, approachable, and easy to use.